Model Predictive Control of Satellite Clusters

Abstract

In this thesis attitude control and relative position control of a satellite cluster is studied. A general introduction to the subject is given, followed by a more detailed study of attitude and relative position control. Two MPC-controllers are derived, one for attitude control and one for relative position control. The performance of the MPC's are compared to the performance of other controllers, derived in earlier theses. For relative position control the MPC is compared to a PID-controller, and for attitude control the MPC is compared to a Lyapunov controller. A Kalman filter is also introduced for estimating the attitude, and analyses of the Kalman filter are done based on the use of two different sensors, a magnetometer and a star sensor. The robustness of the Kalman filter is also studied, and how the loss of measurements from the sensors affects the estimate.